📣 Introducing NVIDIA AI Dev Weekly. 📰 Compiled every Friday, we're delivering technical deep dives, trainings, events, and key updates for developers — all in one read. Stay up to date on the latest in the world of generative AI, data science, accelerated computing, and more👇.

📣 Announcing Llama Nemotron Super v1.5 📣 . This release pushes the boundaries of reasoning model capabilities at the weight class of the model and is ready to power agentic applications from individual developers, all the way to enterprise applications. 📈 The Llama Nemotron

✨ Ready to continuously improve your #agenticAI systems?. 🖥️ Watch to learn how to continuously optimize AI agents for latency and cost — while maintaining accuracy targets. ➡️ Get started at

Your 3D Gaussian Splats look great, but physics? They've been a flop. Discover how to do deformable simulations on 3DGS at #NVIDIAResearch's NVIDIA Kaolin & Warp workshop at #SIGGRAPH2025 👉

🏎️ Just one line of code can accelerate FLUX-1 Dev by up to 2.4X. ⚡ Using FP8 quantization and its seamless LoRA support, Torch-TensorRT unlocks peak GPU performance with maximum flexibility. Technical deep dive:

NVIDIA Dynamo Developer Office Hours

Want to know the secret to making AI smarter?. 🔍 Discover how Chain of Thought reasoning and test-time scaling are revolutionizing AI decision-making. Learn more ➡️

🌄 Scientists used #AI to reveal tens of thousands of previously undetected earthquakes beneath Yellowstone. The discovery shows swarming seismic activity across immature fault lines—transforming how we see the park’s explosive underworld. ➡️ Prior to

RT @nvidianewsroom: This didn’t happen overnight. For over 30 years, NVIDIA has been quietly building the foundation of the AI era — with….

Got a spare weekend? Discover how you can build and train your own reasoning language model in just 48 hours with NVIDIA #NeMo with a single GPU. 🔎 Exploring Llama Nemotron post-training dataset with 32 million samples across areas such as math, coding, chat and sciences.🛠️

Or watch on YouTube:

💬✨Canary-Qwen-2.5B, the new frontier in speech AI. ⚡ This powerful model combines ASR and LLM capabilities, achieving a record-breaking 5.63% word error rate while running 400x faster than real-time on a single A100 GPU for large-scale batch transcriptions. 📺 Hear how this

📐 Celebrating 7² + 24² = 25² . DYK our office buildings are constructed with triangles, as that is the closest approximation to a pixel?. How many 📐 triangles can you count in this view inside our Endeavor building?

🚒 Toxic plumes from industrial accidents, wildfires, and chemical spills spread hazardous materials through urban environments, posing immediate risks to public health and safety. Existing prediction models are too slow for real-time emergency response. Lawrence Livermore

👀 No-code & low-code solutions to accelerate pandas code in Google Colab notebook.

Join us for a session on the basics of NVIDIA Dynamo and disaggregated serving, followed by open Q&A — perfect if you're just getting started or want to learn more about LLM Inference. Add to calendar:

Join us at @CrusoeAI HQ tonight. Details below 👇 to request your spot.

RT @_akhaliq: Nvidia presents ThinkAct. Vision-Language-Action Reasoning via Reinforced Visual Latent Planning